Tirole financial crises, liquidity, and the international monetary system (2002)

Instead of inducing onshore capital to ow o shore to earn higherreturns, these removals have enhanced the appeal of borrowing countries to foreigninvestors by signaling the governments’

FINANCIAL CRISES, LIQUIDITY, AND THE INTERNATIONAL MONETARY SYSTEM

This book is based on the Paolo Ba Lecture given by the author at the Bank of Italy inOctober 2000 The Paolo Baffi Lecture is sponsored by the Bank of Italy

FINANCIAL CRISES, LIQUIDITY, AND THE INTERNATIONAL MONETARY SYSTEM

Jean Tirole

PRINCETON UNIVERSITY PRESS PRINCETON AND OXFORD

Copyright © 2002 by Princeton University Press

Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540

In the United Kingdom: Princeton University Press, 3 Market Place, Woodstock, Oxfordshire OX20 1SY All Rights Reserved

Library of Congress Cataloging-in-Publication Data applied for.

Tirole, Jean

Financial Crises, Liquidity and the International Monetary System / Jean Tirole

p cm.

Includes bibliographical references and index.

ISBN 0-691-09985-5 (alk paper)

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

This book has been composed in Sabon

www.pup.princeton.edu

10 9 8 7 6 5 4 3 2

IMF reforms, regulatory changes, and private sector innovations 18

Institutional and policy responses to market failure 50

5 Identification of Market Failure: Are Debtor Countries Ordinary Borrowers? 77

Is there a need for an international lender of last resort? 110

7 Institutional Implications: What Role for the IMF? 113

Giving the sixth Paolo Ba Lecture on Money and Finance is a great privilege andhonor for me When Albert Ando, on behalf of the scienti c committee, Governor Fazioand the Bank of Italy, asked me to give the lecture, I was both thrilled and intimidated

by the challenge The distinguished lists of economists who preceded me and the Bank’slong-standing tradition of excellence in economic research (a tradition that Governor

Ba helped setting up and that is certainly alive today) provided both high-poweredincentives and anxiety

I could not have written this lecture without the key input of Bengt Holmström (whoco-authored with me a series of papers on aggregate liquidity) and Olivier Blanchard

My discussant, Richard Portes, Ricardo Caballero, Paola Caselli, Mathias Dewatripont,Philippe Martin, Larry Summers, Daniele Terlizzese, and especially Curzio Giannini andOlivier Jeanne gave very detailed and useful reactions to a rst draft in the fall of 2000

I also thank three reviewers for helpful comments

Finally, I would like to thank the Bank of Italy for its remarkable hospitality and formaking the preparation of this manuscript a real pleasure

A wide consensus had emerged among economists Capital account liberalization –allowing capital to ow freely in and out of countries without restrictions – wasunambiguously good Good for the debtor countries, good for the world economy Thetwofold case for capital mobility is relatively straightforward: First, capital mobilitycreates superior insurance opportunities and promotes an e cient allocation ofinvestment and consumption Capital mobility allows households and rms to insureagainst country-speci c shocks in worldwide markets; households can thereby smooththeir consumption and rms better manage their risks Business cycles are dampened,improved liquidity management boosts investment and promotes growth Second,besides insurance, capital mobility also permits the transfer of savings from low- tohigh-return countries This transfer raises worldwide growth and further gives a chance

to the labor force of low-income countries to live better In these two respects, theincrease in the ow of private capital from industrial to developing countries from $174billion in the 1980s to $1.3 trillion during the 1990s1 should be considered good news

That consensus has been shattered lately A number of capital account liberalizationshave been followed by spectacular foreign exchange and banking crises.2 The pasttwenty years have witnessed large scale crises such as those in Latin America (early1980s), Scandinavia (early 1990s), Mexico (1994), Thailand, Indonesia, and SouthKorea (1997), Russia (1998), Brazil (1998–9) and Argentina (2001), as well as manysmaller episodes The crises have imposed substantial welfare losses on hundreds ofmillions of people in those countries

Economists, as we will discuss later, still strongly favor some form of capital mobilitybut are currently widely divided about the interpretation of the crises and especiallytheir implications for capital controls and the governance of the international nancialsystem Are such crises just an undesirable, but unavoidable byproduct of an otherwisedesirable full capital account liberalization? Should the world evolve either to thecorporate model where workouts are a regular non-crisis event or to the municipal bondmodel where defaults are rare? Would a better sequencing (e.g., liberalization of foreigndirect and portfolio investments and the building of stronger institutions for theprudential supervision of nancial intermediaries before the liberalization of short-termcapital ows) have prevented these episodes? Should temporary or permanentrestrictions on short-term capital ows be imposed? How does this all t with the choice

of an exchange rate regime? Were the crises handled properly? And, should ourinternational financial institutions be reformed?

This book was prompted by a questioning of my own understanding of its subject.Several times over recent years I have been swayed by a well-expounded and coherent

proposal only to discover, with striking naivety, that I later found an equally eloquent,but inconsistent, argument just as persuasive While this probably re ected lazythinking on my part, I also came to wonder how it is that economists whom I respectvery highly could agree broadly on the facts and yet disagree strongly on theirimplications.

I also realized that I was missing a “broad picture” An epitome for this lack ofperspective relates to international institutions I have never had a clear view of what,leaving aside the ght against poverty, the International Monetary Fund (IMF) andother international nancial institutions (IFIs) were trying to achieve: avoid nancialcrises, resolve them in an orderly manner, economize on taxpayers’ money, protectforeign investors, respect national sovereignty, limit output volatility, preventcontagion, facilitate a country’s access to funds, promote long-term growth, forcestructural reforms – not to mention the IMF’s traditional current account, internationalreserves and inflation objectives.3

This book is to some extent an attempt to go back to rst principles and to identify aspeci c form of market failure, that will guide our thinking about crisis prevention andinstitutional design Needless to say, I will be focusing on a particular take on theinternational nancial system, which need not exclude other and complementaryapproaches I believe, though, that the speci c angle taken here may prove useful inclarifying the issues

The book is organized as follows Chapter 1 is a concise overview of recent crises andinstitutional moves for the reader with limited familiarity with the topic Chapter 2

summarizes and o ers a critique of economists’ views on the subject Chapter 3 provides

a roadmap for our main argument Basically, I suggest that international nancing issimilar to standard corporate nancing except in two crucial respects, which I name the

“dual-agency problem” and the “common-agency problem” Chapter 4 thereforeprovides the reader with a concise review of those key insights of corporate nance thatare relevant for international nance Chapter 5 describes the market failure Chapter 6

draws its implications for crisis prevention and management Chapter 7 investigates thelessons of the analysis for the design of international nancial institutions Finally,

Chapter 8 summarizes and discusses routes for future research

1 Summers (2000).

2 131 of the 181 IMF member countries have experienced banking problems between 1980 and 1995 (IMF 1996).

3 For example, the Meltzer Commission, or more precisely the International Financial Institution Advisory Commission, chaired by Alan Meltzer and reporting to the US Congress (2000), views the role of the IMF as limiting the incidence of crises, reducing their severity, duration and spillovers.

FINANCIAL CRISES, LIQUIDITY, AND THE INTERNATIONAL MONETARY SYSTEM

Emerging Markets Crises and Policy Responses

Many excellent books and articles have documented the new breed of “twenty- rstcentury” nancial crises.1 I will therefore content myself with a short overview of the

main developments This chapter can be skipped by readers who are familiar withEmerging Markets (EM) crises

The pre-crisis period

No two crises are identical At best we can identify a set of features common to most ifnot all episodes Let us begin with a list of frequent sources of vulnerability in recentcapital-account crises

Size and nature of capital in ows The new breed of crises was preceded by nancial

liberalization and very large capital in ows In particular the removal of controls oncapital out ows (the predominant form of capital control) has led to massive and rapid

in ows of capital Instead of inducing onshore capital to ow o shore to earn higherreturns, these removals have enhanced the appeal of borrowing countries to foreigninvestors by signaling the governments’ willingness to keep the doors unlocked.2

At the aggregate level, the net capital ows to developing countries exceeded $240

bn in 1996 ($265 bn if South Korea is included), six times the number at the beginning

of the decade, and four times the peak reached during the 1978–82 commercial lendingboom.3 Capital in ows represented a substantial fraction of gross domestic product(GDP) in a number of countries: 9.4 percent for Brazil (1992–5), 25.8 percent for Chile(1989–95), 9.3 percent in Korea (1991–5), 45.8 percent in Malaysia (1989–95), 27.1percent in Mexico (1989–94) and 51.5 percent in Thailand (1988–95).4

This growth in foreign investment has been accompanied by a shift in its nature, ashift in lender composition, and a shift in recipients Before the 1980s, medium-termloans issued by syndicates of commercial banks to sovereign states and public sectorentities accounted for a large share of private capital ows to developing countries, andofficial flows to these countries were commensurate with private flows

Today private capital flows dwarf official flows On the recipient side,5 borrowing bythe public sector has shrunk to less than one- fth of total private ows.6 As for thecomposition of private ows, the share of foreign direct investment (FDI) has grownfrom 15 percent in 1990 to 40 percent, and that of global portfolio bond and equityows grew from a negligible level at the beginning of the decade to about 33 percent in

1997 Bank lending has evolved toward short-term, foreign currency denominated debt

Such foreign bank debt, mostly denominated in dollars and with maturity under a year,reached 45 percent of GDP in Thailand, 35 percent in Indonesia and 25 percent inKorea just before the Asian crisis.7

There are several reasons for the sharp increase in the capital ows in the lasttwenty years:8 the ideological shift to free markets and the privatizations in developingcountries; the arrival of supporting infrastructure such as telecommunications andinternational standards on banking supervision and accounting; the regulatory changesthat made it possible for the pension funds, banks, mutual funds, and insurancecompanies of developed countries to invest abroad; the perception of new, high-yieldinvestment opportunities in Emerging-Market economies; and the new expertiseassociated with the development of the Brady bond market.9

Banking fragility Up to the 1970s, balance of payment crises were largely unrelated

to bank failures The banking industry was highly regulated, and banking activity wasmuch more limited and far less risky than it is now It operated mostly at the nationallevel and foreign borrowings were strictly constrained by exchange controls Variousregulations, such as licensing restrictions and interest rate ceilings, kept banks fromcompeting against each other There were also far fewer nancial markets andderivative instruments to play with

The 1970s and 1980s witnessed a trend toward openness and deregulation, but thesubsequent expansion in banking activities and exposure in capital markets madebanking riskier In response, the Basle Committee on Banking Supervision in the pastseveral years has been involved in instituting new banking regulations, concerningminimum capital standards for credit risk (the Basle Accord in 1988), and riskmanagement (the 1996 Amendment to the Accord to account for market risk on thebanks’ trading book), and is proposing some further reforms

A common feature of the new breed of crises is the fragility of the banking systemprior to the crisis.10 Often, the relaxation of controls on foreign borrowing took placewithout adequate supervision For example, banking problems played a central role inthe Latin American crises of the early 1980s.11 The widespread insolvency of Chileaninstitutions in 1981–4 resulted in the Chilean government guaranteeing all foreign debts

of the Chilean banking system and owning 70 percent of the banking system in 1985.Similarly, the banks of the East Asian countries that su ered crises in 1997 (Thailand,Korea, Indonesia, Malaysia) were very poorly capitalized [More generally,overleverage was not con ned to banks as rms’ balance sheets also deteriorated prior

to the crises For example, leverage doubled in Malaysia and Thailand between 1991and 1996, according to the World Bank (1997).]

Currency and maturity mismatch Some of the domestic debt and virtually all of the

external debt of EM economies is denominated in foreign currency, with very littlehedging of exchange rate risk, a phenomenon labeled “liability dollarization” by Calvo(1998) For example, before the Asian Crisis, Thailand, Korea, and Indonesia createdincentives to borrow abroad through implicit and explicit guarantees and other policy-induced incentives.12 To be certain, banking regulations usually mandate currencymatching, but such regulations have often been weakly enforced Furthermore, even if

the banks’ books are formally matched, they may be subject to a substantial foreignexchange risk through their non-bank borrowers’ risk of default For example, theIndonesian private sector engaged heavily in liability dollarization, and so the banksfaced an important “credit risk” (de facto a foreign exchange risk) with those borrowerswho had borrowed in foreign currencies.

The second type of mismatch was on the maturity side For instance, 60 percent ofthe $380 bn of international bank debt outstanding in Asia at the end of 1997 hadmaturity of less than one year.13 Often, the short-term bias has been viewed favorablyand even encouraged by policymakers Mexico increased its resort to de facto short-term(dollar-denominated) government debt, the Tesobonos, before the 1995 crisis SouthKorea favored short-term borrowings and discriminated against long-term capital

in ows Thailand mortgaged all of its government reserves on forward markets Asdocumented by Detragiache–Spilimbergo (2001), short debt maturities increase theprobability of debt crises, although the causality may, as they argue, ow in the reversedirection (more fragile countries may be forced to borrow at shorter maturities)

Macroeconomic evolution Despite attempts at sterilizing capital in ows14 in manycountries, aggregate demand and asset prices grew Real estate prices went upsubstantially

In contrast with earlier crises, which had usually been preceded by large scal

de cits, the new ones o ered more variation in scal matters While some countries(such as Brazil and Russia) did incur large scal de cits, many others, including theAsian countries, had no or small fiscal deficits

Poor institutional infrastructure Many crisis countries have been marred by poor

governance, low investor protection, connected lending, ine cient bankruptcy lawsand enforcement, lack of transparency, and poor application of accounting standards

Currency regime As Stan Fischer (testifying to the Meltzer Commission, 2000) notes,

all countries that have lately su ered major international crises had xed exchangerates (or crawling pegs in the case of Indonesia and Korea)

Summers (2000) usefully summarizes the major sources of vulnerability in recentmajor capital-account crises As Table 1 shows, traditional determinants of exchange-rate crises (current-account and scal de cits) played a role in only some economies Incontrast, banking weaknesses and a short debt maturity seem to have been present inmost of the crises

The crisis

Crises are usually characterized by the following features (in no particular chronologicalorder):

Sudden reversals in net private capital ows Large reversals of capital ows in a short

time interval had a substantial impact on the economies The reversal reached 12percent and 6 percent of GDP in Mexico in 1981–3 and 1993–5, respectively, 20 percent

in Argentina in 1982–3, and 7 percent in Chile in 1981–3.15 In Indonesia, Korea,

Malaysia, Philippines, and Thailand, the combined di erence between the 1997outflows and 1996 inflows equaled $85 bn, or about 10 percent of these countries’ GDPs.

Table 1

Sources of vulnerabilities in recent major capital-accounta

a Notes: Key to table entries: 1, very serious; 2, serious; 3, not central.

b Indonesia let its exchange rate oat in August 1998, did exhibit strong signs of real exchange-rate misalignment, and did not expend reserves defending the rate However, the in exible exchange-rate regime does seem to have encouraged a large buildup of foreign currency debt in the private sector (Source: Summers 2000).

Exchange rate depreciation/devaluation Most countries su ering a crisis were

countries with well-integrated capital accounts and with a xed exchange rate (orcrawling peg) The attacks forced the central banks to abandon the peg or moregenerally to let their currency depreciate Figure 1 illustrates this in the case of Asiancrises For example, South Korea’s won lost half of its value in 1997 Thailand devalued

by 15 percent and after the IMF got involved the baht lost a further 50 percent TheMexican peso lost 50 percent of its value in a week in December 1994 before the IMFintervened The exchange rate depreciation reduced incomes and spending

Figure 1 Asian exchange rate changes, 1997 US dollar per currency, percentage

change, 1 January–31 December (Source Christoffersen–Errunga 2000)

Activity and asset prices Bank restructuring proved very costly.16 Fiscal costs

associated with bank restructurings averaged 10 percent of GDP and have reached muchhigher values Furthermore, whether banks were liquidated or just put on a tighter leash(which was the case for 40 percent of asset holdings in the case of Korean, Malaysian,and Thai banks), restructuring resulted in a credit crunch, which, combined with therms’ own di culties, led to severe recessions, in particular in the non-tradable goodssector Indeed, in Indonesia, Korea, and Thailand, many banks in 1998 not only stoppedissuing new loans, but also cut back on trade credit and working capital.

Equity (see, e.g., Figure 2 for Asian countries) and real estate prices tumbled Asstressed by Krugman (1998), the fall in prices resulted in a wave of inward directinvestment just after massive ights of short-term capital out of those countries Forexample, FDI at re sale prices occurred in South Korea, whose currency had lost half ofits value relative to the dollar, and whose stock market had lost 40 percent of its value

in domestic currency This wave of re sale FDI in some instances (e.g., in Malaysia)gave rise to political concerns of colonization or recolonization

Contagion Some recent crises raised serious concerns about contagion Contagion

occurred in Europe in the ERM crises of 1992–3, in Latin America following the 1994–5Mexican problems (the Tequila crisis), and in Asia in 1997–8 starting with the crisis inThailand (the Asian u) While spillovers have been mostly regional, there are alsoindications that they can be more widespread For example, the August 1998 Russiancrisis spread to Brazil in the fall, triggering the January 1999 crisis, and startedspreading to other Latin American emerging markets Even though the fundamentals inBrazil were weak (large public de cits and uncertainty about the government’s ability

to roll over the public debt), this episode dramatically illustrates the global nature ofspillovers

Figure 2 Asian stock price changes, 1997 Local currency, percentage change, 1

January–31 December (Source Christoffersen–Errunga 2000)

There are several competing hypotheses for the contagious aspect of crises The

portfolio rebalancing hypothesis states that after losing money in one country foreign

investors have to readjust their positions in other countries For example, when Russianmarkets collapsed, some large portfolio managers faced margin calls and liquidatedtheir positions in Brazil Kaminsky et al (2000) argue and o er evidence that mutualfund managers prefer to sell in markets that are mostly liquid, as they incur smallerlosses in such markets Capital adequacy requirements may force banks to adopt similarstrategies Van Rijckeghem and Weder (2000), noting that western and Japanese bankshad substantial exposures in the four Asian crisis countries (Korea, Indonesia, Malaysia,Thailand), present evidence for the hypothesis that a crisis in a country may spread tocountries with common foreign bank lenders, as in the case of Thailand and maybeMexico and Russia It is unclear, though, why investors would deprive themselves ofvery lucrative arbitrage opportunities by failing to manage their regional risks.

A second hypothesis is the trade links hypothesis, which has two versions In the rst,

a crisis in a country has repercussions on countries that are tightly commercially related.For example, the collapse of the Soviet Union had a non-negligible impact on Finland

In the second, competitive devaluation version, crises lead to substantial devaluationsand increased competition for countries producing similar exports

The third hypothesis relates to the existence of common shocks (rise in interest rates,

increase in the price of oil, perceived change in the international community’swillingness to come to the rescue) Although there is then no systemic e ect so to speak,

the crises exhibit a strong correlation The fourth, and nal, hypothesis is a change in

expectations The wake-up-call story asserts that investors realize the lack of solidity of

certain types of economies or the unwillingness of the IMF to help restructure the debt.17Each of these hypotheses probably has some validity, and current research is activelydisentangling their respective impacts in specific crises.18

Rescue packages The international community, often through the IMF,19 designedrescue packages of an unprecedented scale (see Table 2) The 1995 Mexican rescuepackage involved $50 bn or 18 times the country’s quota (while IMF loans havetraditionally been limited to three times a country’s quota),20 and similar size packageswere o ered in Asia in 1997: $57 bn in Korea, $40 bn in Indonesia, and $17.2 bn inThailand It should be borne in mind, though, that despite their huge size, such packages

by themselves were unlikely to restrain speculative attacks on the currencies Forexample, IMF packages in Thailand, South Korea, and Indonesia were much smallerthan the countries’ short-term foreign liabilities Besides, even IMF packages that wouldhave been as large as the countries’ short-term liabilities might not have been su cient

to prevent the crises; Jeanne and Wyplosz (2001) present some evidence that capitalout ows were typically larger than the decrease in short-term liabilities during thecrises.21

Investor bail-in The degree of sharing by foreign investors has been crisis-speci c.

Under the Brady plan (debt writedowns for Latin American countries), creditors got third of the face value of their outstanding claim Investors cashed out at full value inMexico in 1995 They lost up to $350 bn in total in Asia in 1997 and Russia in 1998.22

one-Global solutions have favored bondholders relative to equity investors (foreign directinvestment and equity portfolio investment) Forcing private investors to share the

burden has proved hard in the case of sovereign bonds For example, Eichengreen andRühl (2000), in studying the extent of bail-ins in Ecuador, Pakistan, Romania, andUkraine, conclude that attempts at forcing private investors to share the burden havehad limited success overall, but have been a little more successful where renegotiation-facilitating collective action clauses were appended to the bonds (Pakistan andUkraine).

Table 2 IMF-supported crisis packages of the 1990s: Total financing and outstanding

obligations to IMF (in percent of initial GDPa) (Source: Jeanne-Zettelmeyer 2000)

a GDP in rst year of large package (1997 for Indonesia, Korea and Thailand, 1995 for Mexico and Russia, and 1998 for Brazil)

b IMF disbursements minus repurchases by end-99 related to the program.

c Discounted to the first program year using IMF rate of charge

d Russia had several consecutive IMF programs during the 1990s The rst large-scale program was a stand-by arrangement approved in April of 1995.

e Total Disbursements in the 1990s

A typical debt restructuring proceeds in the following manner: some scal and otheradjustment is demanded from the country while bilateral o cial creditors (the ParisClub) agree to rollover or reschedule some debt claims, and multilateral creditors (theIMF, the World Bank (WB), and other multilateral development banks) bring in newmoney The rest of the external nancing gap is meant to be covered by the privatecreditors through “private sector involvement” (PSI) The level of multilateral support isrelatively well determined IMF and WB lending receives priority The claims ofbilateral creditors are junior, and last come private claims Roubini (2000) argues,though, that Paris Club claims are de nitely not senior to private creditors’ claims:unlike the latter, they are not subject to litigation risk or acceleration23 or formaldefault Accordingly, some countries have kept access to nancial markets even thoughthey were in arrears with bilateral official creditors.24

Conditionality Besides the general prohibition of actions such as the introduction of

new exchange restrictions speci ed in the Articles of Agreement, the IMF usuallyimposes further, country-specific conditions

Traditionally, the IMF performs an in-depth analysis of the sources of economicimbalances Until the 1970s and the 1980s, its conditions focused on current accountbalance and its macroeconomic determinants (most notably monetary and scalpolicies) Under sharp criticism concerning its narrow focus, the IMF then addedmedium-term growth.

More recently, and in particular with the Asian crises, the IMF, while pursuing thetraditional current account determinants,25 has added microeconomic programs such as26– the closure of insolvent nancial institutions (Korea, Indonesia, Thailand), and the

recapitalization of others with explicitly limited public funds (all countries)

– the strengthening of prudential regulation (all countries)

– the liberalization of foreign investments in domestic banks (Korea, Indonesia,Thailand)

– the closure of non-viable rms (Korea) and the restructuring of corporate debts(Indonesia, Korea, Thailand)

– the strengthening of the legal infrastructure and enforcement (competition policy,bankruptcy laws, corporate governance, privatization, etc.)

– the reduction of import tariffs and export taxes (Indonesia), and

– the design of social policies to protect low-income groups and the unemployed, andhealth and education programs

Even observers favorable to conditionality, such as Goldstein (2001), have wonderedwhether the IMF was not su ering from a “mission creep” And a number of economists,including Feldstein (1998a,b), have advocated a return to the old mandate of pursuingmacroeconomic and currency stabilization

A di erent type of criticism leveled at IMF conditionality relates to the programs’credibility IMF policy conditions are often renegotiated, sometimes (as in Asia) within afew weeks of the programs being agreed For example, Indonesia, Korea, and Thailandwere quickly allowed to incur a small budget de cit, and capital adequacy and bankclosure requirements were relaxed for Indonesia and Thailand.27

IMF reforms, regulatory changes, and private sector

– Information collection and surveillance: The IMF has launched a Special Data

Dissemination Standard (SDDS), which provides a checklist of the country’s nancialand economic data In collaboration with the WB, and in consultation withsupervisory agencies, central banks, and the private sector, the IMF collects andanalyzes information published in the reports on the Observance of Standards and

– New forms of lending: The Supplemental Reserve Facility (SRF), created by the IMF in

December 1997 and rst used in Korea, allows the IMF to make large short-term28loans at rates higher than it normally charges SFRs have quickly developed into amajor form of IMF lending In April 1999, the IMF established a no-penalty-rateContingent Credit Line (CCL) to facilitate a rapid disbursement to pre-quali edmembers The drawing of the line is contingent on the IMF’s judgement aboutwhether the country has contributed to its problems The country must apply inadvance for a CCL.29

Besides IMF reforms, experiments are underway that aim at providing privatesolutions to country-level problems While the credit lines involved are relatively smalland therefore very unlikely per se to prevent a crisis, these experiments are worthconsidering The pioneer in the area was Argentina (Mexico, Indonesia, and SouthAfrica have reached similar agreements) Argentina had been badly hurt by the MexicanTequila crisis, with a drop in deposits of the order of 18 percent during a three-monthperiod and a 5 percent drop in GDP On December 20, 1996, the Central Bank agreedwith fourteen international banks on a rm commitment $6.1bn (8 percent of thedeposit base) liquidity option According to the agreement, the Central Bank had theoption to sell domestic assets,30 such as government bonds, to receive US dollars subject

to a repurchase clause The average maturity of the program was three years; theaverage commitment fee, 32.5 basis points; and the interest rate, roughly 2 percentabove LIBOR The credit line was mostly unconditional, as the Central Bank couldexercise the options as long as the country had not defaulted on its international debt.31

The credit line was meant to be a last line of defense to prevent a run on thebanking system Banks were subject to a remunerated liquidity requirement ininternational reserves equal to 20 percent of deposits Adding the Central Bank’s excessinternational reserves (10 percent of the deposit base), the credit line with privatenancial institutions was at the time of the agreement meant to step in only in case of aliquidity shock exceeding 30 percent of the deposit base

As Giannini (2000) points out, however, we should not expect such arrangements to

be a perfect substitute for public money First, and as we have already noted, theamounts involved are relatively limited Second, they must remain proper credit lines Ifsuch credit lines are secured with high-quality collateral and, further, are subject tomargin calls, they do little to enhance a country’s liquidity That is, the credit linesubstitutes for the collateral as a source of liquidity; and margin calls eliminate some ofthe insurance that is the essence of a credit line Third, and importantly, the banksinvolved in the arrangement may wish to hedge their exposure, for example by sellinggovernment securities short Such behavior may undo country risk management, ascountry borrowing is the sum of private and government borrowings from foreigners

Finally, prudential supervisors are changing the rules that regulate the nancialinstitutions’ investments in Emerging Market countries Designing good prudential rules

is in general quite di cult, and particularly so in the case of cross-border investments

For example, the 1988 Basle Accord, which harmonizes capital adequacy requirementsfor banks, requires an equity level of 8 cents per dollar (a risk weight of 100 percent)invested in a loan (with maturity over a year) to a non-OECD bank or sovereign, 0 centsfor an investment in an OECD sovereign bond, and 1.6 cents (risk weight of 20 percent)for a loan to an OECD bank Clearly, the binary criterion “OECD-non OECD” poorlyaccounts for individual situations Ironically, Mexico and Korea became OECD membersjust before their respective crises, which further fueled bank loans to those countries.32

The creation of new derivative instruments and the banks’ ability to take indirectexposures through interactions with hedge funds that are highly exposed to interest rateand exchange rate uctuations (such as Long Term Capital Management during the

1998 Russian crisis) raises new challenges for prudential regulation For example, whilethere is no reason to regulate hedge funds, which in particular are not backed by publicmoney, the banks’ portfolio, credit, and counterparty risks incurred in the interactionwith such funds must be properly assessed by prudential regulators

The new rules proposed in June 1999 by the Basel Committee on BankingSupervision include the use of ratings of sovereign debt in the determination of riskweights, and would leave open the possibility for large banks of using their internalratings (following the socalled ‘internal model approach’) Such ratings would of courseaccelerate ights of capital out of countries that are starting to experience distress Theywould induce banks to scramble for exits (and probably to lend short), on the basis thatadvanced countries’ bank owners are playing with public money and not just theirown.33

1 E.g., Bordo et al (2001), Caballero (2000), Corsetti (1999), De Gregorio et al (1999), Dornbusch (1998), Eichengreen (1999a), Fischer (1998a,b), Hunter et al (1999), Kenen (2000), McKinnon–Pill (1990), Mussa et al (1999), Obsfeld–Rogo (1998), Portes (1999), Rogo (1999), Sachs–Radelet (1995), Sachs–Warner (1995), Summers (1999, 2000), Woo et al (2000), World Bank (1997, 1998), World Economic Outlook (1998) Some observers establish a ner distinction between the crises of the 1980s and those of the 1990s Michel Camdessus, former IMF managing director, called the 1994–5 Mexican crisis the rst nancial crisis of the 21st century There is little purpose in engaging in such a distinction given the limited purpose of this chapter.

2 For such a signal to be credible, though, a government that is committed to capital-account liberalization must nd it less costly to lift controls on capital out ows than a government that is not committed See Bartolini–Drazen (1997) for a formalization of this idea.

3 World Bank (1997).

4 World Bank (1997).

5 See Gourinchas et al (1999) for evidence on lending booms Among other things, this paper suggests that lending booms are not damaging to the economy, although they substantially increase the probability of a banking or balance of payment crisis Also, the proportion of short-term debt rises with investment during the build-up phase.

6 World Bank (1997).

7 The Economist (1999).

8 See De Gregorio et al (1999) and The Economist (1999) for a lengthier discussion of the sharp increase in capital ows from developed countries to developing countries We should note, though, that despite this sharp increase it is still the case that a small amount of capital ows from rich to poor countries Kraay et al (2000) present useful evidence on

“country portfolios” Based on a sample of 68 countries, accounting for over 90 percent of world production, from 1966 through 1997, they show among other things that countries hold small gross asset positions and that these assets are mainly loans For example, industrial countries hold about 3.3 percent and 3.9 percent of their wealth in foreign equity assets and liabilities These proportions are about 11 percent for foreign loan assets and liabilities Relatedly, it is well

known that individuals hardly hedge risks across countries Over 90 percent of US and Japanese nancial portfolios (and

89 percent and 85 percent of French and German portfolios) are invested in domestic assets (French–Poterba 1991), which furthermore are positively correlated with the individuals’ non-financial wealth (human capital) It is also well-known that consumption is less correlated across countries than output, in contrast to what portfolio diversi cation would suggest See Lewis (1999) for a thorough survey of the home bias in equities and consumption.

9 Calvo (1998, 2000) argues that the securitization of non-performing sovereign debt under the Brady plan forced nancial institutions to learn about the economies’ fundamentals and made them more willing to buy securities in the corresponding countries.

10 This fact is well documented by Kaminsky and Reinhart (1999) See also Goldfajn–Valdes (1997) for an analysis of Chile, Finland, Mexico and Sweden.

11 See, e.g., Diaz–Alejandro (1985) and Harberger (1985).

12 For example, Thailand o ered tax breaks on o shore foreign borrowing In contrast, Taiwan had well-capitalized banks with little currency and maturity mismatches Despite a contagious attack on its currency, which forced o cials to float the rate, the Taiwanese economy suffered little from the 1997 crisis.

13 The Economist (1999).

14 Remember that a non-sterilized intervention is similar to an open market operation except that the assets purchased are foreign assets rather than domestic ones; it therefore impacts the domestic monetary base To avoid a ecting the domestic monetary base, the Central Bank can engage in an o setting domestic intervention by selling domestic bonds Thus, in reduced form, a sterilized intervention amounts to purchasing foreign assets by selling domestic ones (or the reverse).

15 World Bank (1997, Figure 1.5).

16 Estimates provided by Rojas-Suarez–Weisbrod (1996) put the bill for bank restructuring at 19.6 percent of GDP for Chile and 13 percent for Argentina in the early 1980s, and from 4.5 percent to 8.2 percent of GDP for the Scandinavian countries in the late 1980s-early 1990s Caprio–Monahan (1999) estimate an average cost of government bailouts in a sample of 59 banking crashes to be 9 percent of GDP in developing countries and 4 percent of GDP in industrialized countries See also Frydl (1999)’s survey.

17 Still another hypothesis is that contagion is triggered by the correlation of “sunspots” across countries in situations

of multiple equilibria (Masson 1999a,b) For example, foreign investors in country B view the fact that there is a run on country A as a signal that there will be a run in country B and engage in a self-fulfilling run.

18 See, e.g., Chang–Majnoni (2000), Corsetti et al (2000), De Gregorio–Valdes (2000), Dornbusch et al (2000), Goldfajn– Baig (2000), Kaminsky et al (2000) and Van Rijckeghem–Weder (2000).

19 The IMF’s role as a crisis manager has expanded over the last few years Although present, the Fund’s crisis management mission was certainly not emphasized in the 1944 Articles of Agreement: “The purposes of the International Monetary Fund are:

(i) To promote international monetary cooperation through a permanent institution which provides the machinery for consultation and collaboration on international monetary problems.

(ii) To facilitate the expansion and balanced growth of international trade, and to contribute thereby to the promotion and maintenance of high levels of employment and real income and to the development of the productive resources of all members as primary objectives of economic policy.

(iii) To promote exchange stability, to maintain orderly exchange arrangements among members, and to avoid competitive exchange depreciation.

(iv) To assist in the establishment of a multilateral system of payments in respect of current transactions between members and in the elimination of foreign exchange restrictions which hamper the growth of world trade.

(v) To give con dence to members by making the general resources of the Fund temporarily available to them under adequate safeguards, thus providing them with opportunity to correct maladjustments in their balance of payments without resorting to measures destructive of national or international prosperity.

(vi) In accordance with the above, to shorten the duration and lessen the degree of disequilibrium in the international balances of payments of members The Fund shall be guided in all its policies and decisions by the purposes set forth in this Article.” (Article I).

20 Total rescue packages can exceed the IMF’s treble-quota limit as a) they may involve bilateral creditors or other IFIs, and b) the IMF limit may be exceeded in case of contagious (systemic) impact.

21 See also Bernard–Bisignano (2000) Relatedly, Haldane (1999), reviewing the evidence, concludes that IMF rescue packages fail to reassure investors.

22 This number is given by Hildebrand in De Gregorio et al (1999, p118).

23 “Acceleration” refers to the possibility for a category of debtholders to demand early reimbursement if there is

default on another claim.

24 See Roubini (2000) for a broad discussion of PSI, including the issue of which claims ought to be included (should PSI include bonded debt, short-term interbank loans, Euro bonds, domestic debts?) and that of whether PSI should be accompanied by exchange rate and capital controls.

25 In this respect, demands such as improvements in tax collection (Russia) and the reduction in local spending (Brazil) are traditional ones.

26 See World Economic Outlook (1998, p105) for a broader list.

27 See Radelet–Sachs (1998) for more detail Further thoughts about IMF programs can be found in Dixit (2000), Kaufman–Kruger (2000), Hunter (1999), Masson (1999), Mussa–Savastano (1999)

28 For longer-term (say, over three years) adjustments of macroeconomic imbalances, the IMF can use di erent programs: the Extended Fund Facility (EFF), and the Enhanced Structural Adjustment Facility (ESAF, at low interest rates for low-income countries).

29 The Meltzer Commission report of February 2000 argues that, together with the existence of other channels for IMF money, the application requirement explains why no country had yet applied at the date of the report The commission’s argument is that an application would be interpreted as a signal of an impending crisis In September 2000, the IMF tried

to enhance the appeal of the Contingent Credit Line by getting rid of the commitment fee, by reducing the interest rate penalty, and by relaxing conditions for prequalifìcation.

30 There was overcollateralization: Argentinian bonds had to exceed by 25 percent the value of funds delivered WB/IDB resources are used to cover further margin calls, making the contingent liquidity facility not a purely private arrangement.

31 The Mexican government has arranged an overdraft facility for about $3 bn Amusingly, the government’s decision to draw $2.66 bn in September 1998 aroused much controversy among lending banks (the market interest rates had gone up since the writing of the arrangement).

32 According to the Bank for International Settlements, bank loans to developing countries totaled $931 bn, of which

$520 bn was in short-term loans in December 1997.

33 Monfort–Mulder (2000) estimate that the banks’ capital requirements corresponding to their exposure on Emerging Market lending would have increased by 40 percent under the proposed modi cation to the Basle Accord They also question the relevance of the sovereign ratings provided by rating agencies.

The Economists’ Views

Many of the best minds among economists and the nancial community have expressedtheir views on recent international nancial crises and the design of a new nancialinfrastructure.1 While there is widespread agreement on what happened, there is muchless convergence on what should be done about it Still, we can identify a common core

of proposals (together with, as usual, a few dissenting voices), as well as a number ofissues on which economists disagree Abusing terminology, let us call the former the

“consensus view”

Consensus view

The seven pillars of the consensus view

Most recommendations concur on a number of desirable steps:

• Elimination of currency mismatches A high level of indebtedness in foreign currencies

makes a country very vulnerable to a depreciation in the exchange rate and to theconcomitant liquidity and solvency risk faced by domestic banks and rms Along withthis, the absence of countrywide risk management confronts monetary policy with anunpalatable dilemma.2 A tight monetary policy, to maintain the exchange rate, runs therisk of a severe recession, while a loose monetary policy leads to depreciation of thecurrency and possibly the bankruptcy of rms and banks that are highly indebted inforeign currency

A common proposal, therefore, is to eliminate currency mismatches, at least at thelevel of banks and the government Furthermore, many suggest that a domestic buildup

of international reserves would reassure foreign investors about the value of theirinvestment

• Elimination of maturity mismatches To prevent hot money from eeing the country,

many advocate a lengthening in debt maturity, as well as measures encouragingalternatives to short-term debt, such as foreign direct investment (FDI) and investment

by foreign bank subsidiaries

• Better institutional infrastructure In response to the poor governance that has marred

many crisis countries, the consensus view argues that infrastructure-promoting reforms,such as adherence to universal principles for securities market regulation designed bythe International Organization of Securities Commission (IOSCO) and those foraccounting designed by the International Accounting Standards Committee (IASC),

would reassure foreign investors and help prevent crises.

• Better prudential supervision Most crisis countries’ prudential regulations satis ed the

international standards as de ned by the Basle Accord (1988, revised in 1996) It is inthe nature of such standards to be highly imperfect (despite much e ort devoted to theirdesign by the Basle Committee on Banking Supervision) and to leave substantialdiscretion with the national supervisory authorities Indeed, enforcement of thestandards in a number of crisis countries has been highly negligent, resulting in lowcapital adequacy and high values at risk The consensus view calls for a betterenforcement of existing prudential regulations

• Country-level transparency Most economists recommend that foreign investors be

informed in a uniform and regular manner of the country’s structure of guaranteed debtand off-balance-sheet liabilities

• Bail-ins There is widespread agreement on the desirability (although not on the

feasibility) of forcing the foreign investors to share the burden in a case of crisis.3 Theargument is that bailing-in the investors will force them to act in a more responsiblemanner in lending only to countries with good fundamentals

There is disagreement fundamentals about whether bail-ins are indeed feasible forshort-term debt claims (in the absence of mandated rollover) While there is a consensusaround the view that bail-ins of bondholders may be facilitated by the existence ofcollective action clauses (in the same way that investors had to share the losses on thesyndicated bank loans of the 1970s), De Gregorio et al (1999) argue that the bail-inpolicy is not time-consistent as international nancial institutions (IFIs) are unlikely tostay passive and let the crisis unfold when bondholders refuse to write down some oftheir claims Eichengreen and Rühl (2000) concur and add that even if bail-ins aresuccessful in a given country, they may induce investors to readjust their portfoliostoward large countries, with systemic implications, in which bail-ins are less likely

• Avoid xed exchange rates The reader will probably be surprised that, in this

discussion of economists’ consensus views on international crises, I kept views onexchange rate regimes for later There are two reasons for this First, many of the recentcrises seem to have been triggered by fundamentals somewhat unrelated to exchangerate regimes The exchange rate regime, however, has an important impact on crisisresolution and its consequences Second, there is both a consensus and con ictingadvice The broad consensus4 is that xed exchange rates work poorly under nancialderegulation and that countries with open capital account should choose betweenfloating rates and hard pegs.5

Burnside et al (1999, 2001) suggest abandoning xed rates as soon as possible, andthe Meltzer Commission (2000) and Sachs–Woo (2000) recommend avoiding pegged oradjustable rate systems The options favored currently are currency boards and unions(Dornbusch) and oating exchange rates (Eichengreen), even though many countriesstill prefer to manage their exchange rates It is clear, though, that countrywide crises,

be they triggered by poor domestic policies, a jump in the oil price, a sharp drop in acommodity price, a change in international interest rates, increased trade competition,

or a foreign recession, will still exist regardless of the exchange rate regime

A critique

There is no denying that steps in the direction de ned by the pillars of the consensusview would reduce a country’s risk of crisis and thereby reassure foreign investors Theconsensus view raises some hard questions, though:

• First, that of the objective function: Preventing crises cannot be a goal in itself; afterall, prohibiting foreign borrowing would eliminate the threat of foreign debt crisisaltogether! The question therefore is, how desirable are those policies when trying toaccomplish a well-stated, unambiguous objective? This leads one to question the rsttwo pillars concerning dangerous forms of nance There is a nagging suspicion that

one may be addressing the symptoms and ignoring the fundamentals As Jeanne (2000)

argues with regards to currency and maturity mismatches,

It is di cult to assess the relative merits of these reforms, however, without understanding the underlying reasons why mismatches arise in the balance sheet of emerging countries.

The reform proposals aimed at eliminating dangerous forms of nance may well bemisleading if they are based solely on the consideration of the ex post (crisis time) e ect

of mismatches and ignore both their private rationale and their social impact on ex ante(pre-crisis) government incentives

For example, as argued by Jeanne (1998, 2000) and Kashyap (1999), short-term debt

is a natural reaction to the uncertainty faced by foreign creditors with respect toeventual repayment.6 Keeping an exit option is a standard protection for creditorsagainst abuse by the debtor Jeanne (2000) builds an interesting model of the maturitymismatch in which the bene t of short-term debt over long-term debt is that the threat

of a fundamental-based run induces the government to implement a scal adjustment7.The cost is the possibility of a non-fundamental-based run He studies the impact ofvarious proposals in the context of his model and shows, for example, that taxing short-term debt ows is unambiguously welfare-decreasing This illustrates the importance oflooking also at fundamentals and not simply at symptoms

The possibility that we are treating the symptoms suggests that we ought to return tofirst principles

• Second, and assuming that they are indeed desirable, one may ask, why were thosegreat policies not adopted earlier, and if some obstacle has to be removed, how will theycome about in the future? That is, the consensus view must address the question of whythe sure- re recipes failed to be implemented In this respect, it faces a di cult choicebetween two hypotheses:

– Government incompetence According to this hypothesis, the proposed reforms were

never put in place because the country’s politicians and bureaucrats were utterlyincompetent While this has probably been the case in some instances, I would nottake this hypothesis too seriously, at least not as universal First, governments oftenpretend not to hear international advice when the latter con icts with their politicalinterest Second, under the incompetence hypothesis, the IMF and other IFIs wouldhave a remarkably straightforward job: their experts would simply have to

evangelize the consensus view among politicians and bureaucrats (perhaps inprivate, as the latter may not want to admit incompetence) The country’s politiciansand bureaucrats would then embrace and obediently follow the advice Thehypothesis in this respect seems counter to what one would expect to happen.

– Political economy According to this hypothesis, the proposed policies were never

implemented because it was not in the interest of the governments to do so While

this is my favorite hypothesis of the two, it is not, as it stands, without di culties It suggests that, to have any content, such policies should be imposed by foreigners

upon a country and its legitimately elected government Such a paternalistic attitudecon icts with much of the economic experts’ proper concern about countryownership and the widespread criticisms of IMF arrogance

Conflicting advice and the topsy-turvy principle

While reaching a consensus view on the seven pillars, economists disagree on manyother issues Let me help explain why they do [and why they may keep on disagreeing

until an empirical consensus is reached] At an abstract level, the conflicting advice stems from a key trade-o between ex ante (pre crisis) incentives and ex post e ciency

(satisfactory crisis resolution) Orderly workouts may con ict with the incentives of

countries and investors to avoid crises In other words, con dence is often the mirrorimage of moral hazard

At the risk of oversimplifying, for each issue I shall divide economic experts intothose who stress ex ante borrower and lender incentives, and those who aim at e cientcrises resolutions (individual positions vary with the issue) The former (let’s call them

“hawks” for convenience) focus on deterrence and recommend tough sanctions againstcountries that misbehave and international investors who overlend Arguing that criseswon’t fail to occur anyway, the latter (the “doves”) are more concerned about anorderly workout of the crises

The topsy-turvy principle states that for policies of interest, there is an inherentcon ict between ex ante incentives and ex post e ciency There is nothing mysterious

in this con ict, which is faced by ordinary corporations when designing a nancingscheme and a corporate governance framework Let us illustrate it through a few issues

• IMF liquidity provision The IMF has moved in the direction of a lender of last resort

(LOLR) by adopting, in 1997, an emergency nancing mechanism to speeddisbursement, and by extending its ability to lend into arrears (that is, to lend to acountry that is in arrears to private creditors and has not yet renegotiated these privatedebts).8 This move is consistent with the advice of the many economists who are dovish

on this issue.9

The hawks, e.g Schwartz (1999), resolutely oppose the LOLR function on the basisthat it encourages countries (or investors) to overborrow (or overlend) Others formallyendorse the LOLR function, but qualify this endorsement through various restrictions

For example, the Meltzer Commission (2000) recommends that IMF loans have a shortmaturity (typically a few months), pay a penalty rate, be senior to all debts and besecured by government bonds There are, of course, limits as to what such loans canachieve As is the case for banks, for which discount window lending at penalty rates ordeposit insurance premia indexed on capital adequacy would aggravate di culties andexacerbate moral hazard,10 penalty-rate emergency loans, while providing better exante incentives than penalty-free emergency loans, also con ict with an ex post

e cient resolution of the crisis Furthermore, liquidity loans, almost by de nition,cannot be fully secured Real assets that are used to secure the liquidity loans, therebymaking them almost riskless, cannot be employed to attract alternative forms offinancing In other words, secured liquidity loans crowd out other forms of liquidity

• IMF conditionality The concept of conditionality has recently been extended to

encompass a broad array of macroeconomic policies and institutions

The doves in the matter, a group of critics on the left and right, argue that IMFpolicies are too intrusive (Feldstein 199811), do not re ect country ownership, and hurtthe poor (Meltzer Commission 2000) or are unnecessarily constraining (Stiglitz).12

• Capital controls Economists’ conventional wisdom in the matter of capital controls has

until recently been that nancial integration is bene cial to the host country and thatcapital controls, besides not always being e ective, are welfare-reducing The primaryargument behind this broad consensus is that the host country can tap internationalsources of nance to fund its development and to diversify the country risk It is alsobeen argued that foreign investment promotes the development of domestic nancialmarkets and that FDI brings know-how to the domestic industry

This consensus has been shattered lately While Summers (2000) calls for theelimination of biaises13 that lead to an excessive accumulation of short-term debt andargues that controls of short-term ows tend to be ine ective over time and discouragethe integration of nancial services that can be an important source of stability, severalother prominent economists have advocated direct or indirect restrictions on capitalows Baghwati (1998) argues that the bene ts from nancial integration apply only tolong-term direct investment Krugman (1998) proposes curfews on capital out ows.Proponents of controls on short-term capital in ows include Sachs–Woo (2000),Eichengreen (1999, in favor of Chilean-style14 taxes on short-term capital in ows), andCaballero (2000, who argues that taxing capital in ows contingently, i.e., removing thetaxes during external distress, may be justified).15

Leaving aside the technical question of whether capital controls are indeed e ective

or can be evaded through various schemes,16 as well as the possibility that capitalcontrols may engender corruption, most would agree that taxing short-term foreigninvestments both reduces ex post the probability of a crisis and makes it harder ex antefor the country to access capital, as there is a reason why investors are unwilling to lendlong in the first place

A similar point applies to the proposal of IMF-sanctioned standstill, or to proposalsencouraging suspensions of convertibility.17 A couple of recent proposals18 have pushedfor “Chapter 11”-style petitions for debt relief.19

• Orderly workouts The doves in the matter include Eichengreen and Portes (1995,

1997, see also Portes 2000) who have argued in favor of creditor committees.20 Thegeneral idea is to facilitate renegotiation of foreign debt so as to avoid ine cientliquidations in situations of foreign distress Loan contracts would be rewritten in order

to clarify the representation of creditors by introducing a collective representationclause for sovereign bond agreements and to permit a quali ed majority vote torestructure lending terms.21 The orderly workout would then be enabled by the IMFlending into arrears, that is, providing nance to the country even when the latter is inarrears to private creditors

As is well-known, bond issues are as a rule harder to renegotiate than commercialsyndicated bank loans, which involve many fewer creditors Many sovereign bondeddebt agreements are written under State of New York law.22 Renegotiation is notfacilitated by this statute, which does not include contractual provisions for quali edmajorities to modify the terms of the bonds and to impose such modi cations onminority bondholders The unanimity requirement provides incentives for minorities todisrupt the renegotiation process to enable them to obtain a favorable settlement fromthe debtor or to be bought out by other creditors Bonds governed by UK law, bycontrast, include provisions that make it easy to call a bondholders meeting andempower a quali ed majority at the meeting to make deals that are binding on theminority as well

Doves in the matter also include proponents of an IMF acting as a coordinator ofcrisis resolution (e.g., Feldstein 1998, Fischer 1999)

Hawks in the matter include investment professionals who are worried that a greaterability to renegotiate debts in a distress contingency makes it more likely that thiscontingency will arise in the future.23 Roubini (2000) studies recent bond restructurings

in Pakistan, Ukraine, Russia, and Ecuador; these restructurings occurred through debtexchange o ers and, in the case of Russia and Ecuador, there were no collective actionclauses He argues that collective action clauses have played a minor role Whilepointing out that exchange o ers are facilitated by the presence of collective actionclauses that bind-in holdouts, Roubini expresses concern about the possibility thatstructured renegotiation would prove too bureaucratic

“Unrealistic” encroachments on sovereignty

Arguing that a country’s governments and institutions are not to be trusted, a number ofeconomists have suggested that key sovereign control rights should be transferred tonew supranational institutions Cooper (1984) proposes the creation of a worldcurrency and Central Bank; Sachs (1995), that of an international bankruptcy court;Kaufman (1998), Krugman, Eatwell, and Taylor, the creation of a world nancialregulator; and Soros, that of an International Central Bank acting as lender of lastresort

While I concur with Eichengreen (1999, 2000, who calls them

“pie-in-the-sky-schemes”) that such reforms are rather unrealistic at this stage, I also feel that we shouldlisten to visionaries, as some previously unthinkable transfers of power to supranationalauthorities have occurred in the past Still, such schemes should be analyzed in a broadercontext and not solely as a patch to existing foreign debt crises In particular, theyshould be assessed in the light of recent and future work in political economy on theoptimal size of countries and degree of subsidiarity.

The fundamental view argues that appearances are deceiving In this case, thenancial sector’s low capitalization and poor risk management exposed governments tolarge implicit liabilities When East Asia was exposed to several shocks in the mid-1990s(increasing competition of Chinese exports, recession in Japan, appreciation of thedollar after 1995), leading to further nancial di culties, and foreign investors becameaware of the gravity of the situation, capital flows quickly reversed

The strength of the fundamental perspective is that it builds on well-documentedbanking fragility and implicit guarantees given by governments to domestic and foreigndepositors It makes a coherent argument that the currencies of crisis countries had todepreciate in order to reflect the fundamentals

What the fundamental view implies for the IMF and other IFIs is less clear If weaccept that sovereign authority should be respected, the fact that some governments(e.g., Mexico in 1995, Asian countries), possibly because of crony capitalism, left theircountries with huge tax bills and widespread recessions should not be a motive forintervention, however shocking the outcome After all, the US government’smishandling of the S&L crisis in the 1980s or the French Credit Lyonnais scandal (albeitwith smaller consequences in relative terms) have always been viewed as internalmatters, of concern primarily to domestic voters

It is sometimes argued that things are di erent in the case of EM crises, since so

much money at stake is owned by foreigners This point, however, requires furtherelaboration After all, these foreign investors lend willingly, and are not directly hurt bythe EM taxpayers’ (involuntary) guarantee of bank deposits In contrast, theinternational community may be hurt if the IMF or creditor countries’ governments aredrawn into debtor country bailouts Such an expectation of an international bailout –the “moral hazard play” – seems to have been widespread in the case of Russia in 1998(these expectations failed to be realized, as it turned out) If the creditors and the debtorstall in their negotiation, creditor countries’ governments, to the extent that they enjoylarge gains from trade with the debtor country or have geopolitical stakes in thecountry’s stability, may be drawn into contributing to rescheduling agreements (Bulow–Rogo 1988, 1989) In the presence of such “ex post subsidies”, competitive lenders arewilling to lend more at any given interest rate and the borrowing country is able toextract the corresponding surplus.

The amounts lent by the o cial sector, however, while huge by historical standards,have remained limited relative to the size of recent crises Furthermore, they havereceived preferred creditor status.25 Indeed, there is little evidence that internationaltransfers associated with big international bailouts are commensurate with the domesticinternal transfers Summers (2000) argues that this is unlikely for recent crises Brealey(1999) similarly views the danger of moral hazard created by the prospect of IMFassistance as often overstated:

The subsidy in IMF loans is negligible compared with the losses that have been su ered by investors in East Asia, Russia and Brazil.

Jeanne and Zettelmeyer (2000, 2001) provide some empirical evidence to this e ect.Because the direct losses incurred by o cial lenders in big international bailouts mayunderestimate the true losses (some loans are rolled over), they look at the “mostpessimistic scenario” in which currently existing IMF claims are worthless They ndthat IMF lending subsidies are quite small and that it is the domestic taxpayer, not theglobal taxpayer, who ultimately foots the bill.26

Discussion: What is “moral hazard”? Where is the body? I have eschewed the use of

“moral hazard” even though the “fundamental view” is sometimes also called the “moralhazard view” The problem is that the literature has given di erent meanings to “moralhazard”, and these meanings have sharply di erent policy implications Let us thereforetry to sort out the different interpretations

The rst and most frequent interpretation refers to moral hazard on the part of thehost country’s government’s The fundamental view’s starting point – that thegovernment accumulates substantial implicit liabilities through lax banking supervision– refers to an important form of government moral hazard The granting of implicit orexplicit government guarantees leading to overborrowing by domestic agents isaccordingly often referred to as “debtor moral hazard” in commentaries.27 As we willdiscuss in Chapter 5, there are many other relevant forms of government moral hazard,with different implications

Figure 3 Burden Sharing

In the fundamental view, the domestic taxpayer foots the bill More generally, a keydeterminant of incentives is, as is widely recognized in commentaries, who shares theburden Roughly, there are three possible victims: the domestic taxpayers, the foreigninvestors whose equity value is depreciated or debt claim is in default or renegotiated,and the “o cial sector” (which we de ne here as IFIs plus advanced countries’Treasuries) that can lose money in attempting rescues: see Figure 3

As we have seen, the burden sometimes falls entirely on domestic taxpayers, as inMexico in 1995.28 Most often, however, it is shared between domestic taxpayers andforeign investors The latter su er through the loss of value of their equity portfolios,through default of debt claims, or through renegotiation of these claims via privatesector involvement (PSI) Despite many claims concerning bailouts by the IMF and the

US government,29 external bailouts have been relatively limited, as we discussed earlier.Certainly, the fact that most loans have been repaid in itself does not mean that thelarge-scale rescue attempts were not external bailouts, because such loans may havebeen very risky at the time they were issued In any case, with the ever-escalating size

of official loans, we cannot dismiss the possibility of future external bailouts

The concept of “creditor moral hazard”, although often employed, is more delicate.Roughly, it has two meanings: overlending in the pre-crisis stage, and attempting toshift the burden (extract bailout money from the official sector or the domestic taxpayer)

in the case of a crisis Even assuming that investors invest in expectation that o cial ortaxpayer resources will be available for bailouts, the notion of overlending moral hazard(also called the “moral hazard play” in commentaries) – the market bets on a rescue – is,

as its stands, puzzling Investors are motivated by pro t and react to the incentivesgiven to them Foreign investors may well be very happy to lend to weak banks andrms, knowing that the latter will be rescued by an external or internal bailout.Blaming investors for reacting mechanically to incentives distorted by the host

government and/or the o cial sector is really an attempt by the latter to deny theirresponsibilities.30 The issue of investors’ strategies to shift the burden during a crisis incontrast deserves more attention.

Remark: other fundamentalist theories While the dominant “fundamental view”

emphasizes government bailouts, Caballero and Krishnamurthy (2001b) stress privatesector fundamentals as the culprit for ine cient crises Their starting point is theimportant question of whether and why private domestic borrowers underinsure againstcountry shocks and exchange rate depreciation by contracting dollar-denominated debt.Foreign-currency-denominated debt implies that those rms whose assets are indomestic currency su er grave liquidity shocks when the currrency depreciates.Caballero and Krishnamurthy show that rms have no incentive to underinsure ifdomestic nancial markets are well developed, but that they will contract excessivelyforeign-currency-denominated debt if domestic nancial markets are ine cient.31 Theoutcome is an overreaction to country-speci c shocks in countries with domesticfinancial underdevelopment

“Panic” or “multiple equilibria” view

A number of economists (e.g., Calomiris 1999, Cole–Kehoe 1998, Chang–Velasco 1999,Feldstein 1998, Masson 1999a,b, Obstfeld 2000, Sachs 1998) have used an analogy withbank runs to analyze recent crises Bank runs may arise when short-termcreditors/depositors start refusing to roll over their claims on the banks and demandimmediate payment The bank, which performs a maturity transformation function andhas longer maturities on the asset side, is then forced to liquidate some long-term assets

at a cost, making other creditors/depositors worried that they will not be able to recouptheir investment, thereby inducing them to run for the exit The eventual outcome maywell be socially ine cient In particular, creditors collectively would be better o ifeach (hypothetically) could commit to roll over any claims unless they individuallyreally need the money By analogy, many economists have argued that foreign investorsmay scramble for exits out of the country before others do and that this behavior, whileindividually rational, is collectively irrational and, furthermore, hurts the borrowingcountry:

There is no ‘fundamental’ reason for Asia’s nancial calamity except nancial panic itself… Asia is reeling not from

a crisis of fundamentals, but from a self-ful lling withdrawal of short-term loans, one that is fuelled by each investor’s recognition that all other investors are withdrawing their claims Since short-term debts exceed foreign exchange reserves, it is ‘rational’ for each investor to join the panic (Sachs, 1997).

In this basic story, as in the original bank-run models of Bryant (1980) and

Diamond–Dybvig (1983), the panic outcome is one possible outcome (if other depositors

do not panic, I have no reason to panic) The panic view has therefore been criticized aslacking predictive power.32 Recently, though, richer models have been developed thatmay predict unique outcomes (panic or no panic): see Morris–Shin (1998) and Rochet–Vives (2000)

When compared with the fundamental view, the panic view leads to quite di erent

implications As fundamental imbalances are not the cause of the crisis, policytightening may do more harm (i.e., exacerbate the crisis) than good Rather, the panicview leads to the conclusion that crises should be prevented either by avoidingdangerous forms of debt, or through access to a credit facility, or both Manyinstruments can be used for this purpose:

– capital controls (e.g., controls on capital out ows to eliminate the panic, reserverequirements to limit short-term foreign borrowing)

– elimination of currency and maturity mismatches in foreign borrowing

– IMF-sanctioned standstills (debt service suspensions, suspension of convertibility) ordebt-equity swaps, and

– credit line available either from the IMF or from large private nancial institutions(as in the case of Argentina and Mexico)

With so many instruments available, it is hard to see how a crisis could occur Thepanic view, therefore, is somewhat incomplete as it does not account for the feasibleinstitutional responses to the underlying problem The fact that the “anti-run devices”are not always in place suggests that there are deeper political and economic forces inplay that must be accounted for before we can draw with con dence the policyimplications of the panic view The panic view, as it stands in the narratives, does notaccount well for the existence of crises and therefore does not yet provide a rmfoundation on which to base policy recommendations My “gut feeling”, though, is that

it is well worth pursuing research in this direction Indeed the panic view is useful as asensible-looking building block to describe what may happen in speci c circumstances,and (in those versions of the theory predicting a unique outcome) to analyze the timingand dynamics of the crisis

1 The reader will nd much useful material on the topic in Nouriel Roubini’s remarkable website:

5 Hard pegs refer to xing the exchange rate to a hard currency and holding enough reserves (say, equal to the monetary base) to back up the peg.

6 To be certain, some of the maturity mismatch is policy induced; for example, Korea discriminated against long-term foreign investment But there is no doubt that the main reason for the maturity mismatch is investors’ concerns.

7 In the same way short-term debt constrains misbehavior by an ordinary corporation: see, e.g., Calomiris–Kahn (1991), Jensen (1986), and Rey–Stiglitz (1994).

8 In 1998, the IMF modi ed its 1989 policy allowing the Fund to lend to members after sovereign arrears to commercial

banks have emerged but before agreements to restructure such debts have been reached The 1998 modi cation extends the policy to cover sovereign arrears to other (nonbank) creditors, and to nonsovereign arrears to private creditors arising from the imposition of exchange controls.

9 E.g., Allen–Gale (2000a), Chang (1999), Eichengreen (1999), Fischer (1999a, b), Krugman (1999a), Obstfeld (2000), Sachs (1998).

10 See Dewatripont–Tirole (1994a) for a formal analysis of the incentives provided by such policies.

11 Feldstein argues that the IMF attitude in Asia showed much arrogance and usurped sovereign responsibilities.

12 The IMF has also been severely criticized for its time-inconsistency As already discussed, some of the tough policies imposed on Asian countries (e.g., the fiscal targets) were quickly relaxed.

13 Restrictions on foreign direct investment or on access by nonresidents to long-term bond markets, and tax incentives that favor debt over equity instruments.

14 Chilean-style unremunerated reserve requirements (which were abandoned in Chile in 1998) impose a tax on term capital in ows From 1992 through 1998, 30 percent of in ows had to be deposited at no interest at the Chilean central bank for a one-year period The shorter the maturity, therefore, the larger the tax.

short-15 Similarly, a report from a group sponsored by America’s Council on Foreign Relations (including Bergsten, Eichengreen, Feldstein, Goldstein, Krugman, Soros, and Volcker) recommended the use of a holding tax on short-term capital inflows for countries with weak financial systems.

16 One issue is that domestic agents can enter o shore swap contracts with foreign holders of (what resembles) term debt See Garber (1998) for a good account of possible schemes that can be designed to evade capital controls.

long-17 Remember for instance that convertibility of short-term loans to Korea was suspended in December 1997.

18 E.g., Schwarcz (2000).

19 In reference to the US bankruptcy law Under Chapter 11 bankruptcy, the rm les for bankruptcy but continues to operate under the same management Management designs and proposes a reorganization plan that speci es how claims on the rm are altered or compensated (e.g., debt-equity swaps, debt forgiveness, etc.) The reorganization must be approved

by a majority of shareholders and a majority of each class of creditors.

20 The Eichengreen–Portes proposal has been very in uential in policy circles In 1996, the Rey committee recommended the use of collective action clauses to facilitate the restructuring of bonded debt The proposal was also received favorably in various fora such as the G-7 and by the IMF.

21 Along these lines, the G-10 (1996) suggested including majority voting, sharing of haircuts, and non-acceleration clauses for sovereign bonds The G-22 (1998) took a similar stance on the issue On April 21, 1999, Rubin, then US Secretary of the Treasury, argued strongly in favor of such covenants Eichengreen–Portes (1997) and De Gregorio et al (1999) argue that the potential for negative signaling by countries adopting such provisions can be undone by instructing the Fund to lend at more attractive rates to countries that issue debt securities with such provisions.

22 See the 1999 IMF note on “The IMF Policy on Lending into Arrears to Private Creditors” for more detail.

23 For example in Jeanne’s (2000) theoretical model, the institution of a creditor committee to renegotiate sovereign debt reduces the cost of a crisis and has two e ects: a bene cial one (reduction in the cost of unavoidable runs) and an adverse incentive one Jeanne correctly argues that one should make a distinction between renegotiation-friendly measures, such as the institution of a creditor committee, which facilitates an orderly workout, and borrower-friendly measures, which shift the bargaining power in the renegotiation towards the borrower The adverse incentive e ect disappears if the borrower has little bargaining power While it is straightforward to facilitate renegotiation through institutional design, it

is less easy simultaneously to facilitate renegotiation and create a low bargaining power for the borrower More research on this topic is warranted.

24 Good discussions of the theories can be found, e.g., in Corsetti (1999) and Eichengreen (1999).

25 As Roubini (2000, p26) points out: “The simplistic idea that a sovereign would follow reckless policies that lead to nancial distress for the country in order to end up receiving IMF assistance does not make too much sense First, such policies are highly costly, in terms of output, in ation and other welfare measures; second, sovereigns tend to resist the idea of asking for IMF support as it is perceived to be costly; third, IMF assistance comes with strings, conditionality and is subject to often painful adjustment policies On the other hand, it is also true that, while a sovereign may not purposely follow reckless policies to get IMF support, its policies may at the margin be biased towards risky and unsound behavior if there is some expectation of external nancial support in case of trouble.” The expectation of an external bailout may nevertheless play a role in the case of large countries – with systemic consequences such as in the case of Russia, which was deemed to be “too nuclear to fail”.

26 “Based on the historical record, an EMBIG economy [country belonging to J.P Morgan’s “EMBI Global” Index, a group of 27 relatively advanced emerging countries] that borrows 10 percent of its GDP would fail to repay with a

probability of at most 5 percent – the probability of an “in nite cycle” for this class of countries The resulting implicit transfer – due to the fact that the interest rate charged by the IMF fails to re ect this default risk – is thus less than 0.5 percent of the country’s GDP If the country represents 1 percent of the world population and GDP (this corresponds to a large emerging economy, between Argentina and Brazil in size), the per capita cost of the bailout for the global taxpayer would amount to less than 0.0005 times that borne by the domestic taxpayer It bears emphasizing that even these small numbers are based on an extreme assumption underpinning our hypothetical worst case scenario, namely, that none of the outstanding debt on “in nite” lending cycles will be recovered Thus, a reasonable estimate of the ex ante subsidy implicit

in IMF lending is likely to be much smaller.”

27 See, e.g., Calvo (2000b) and Roubini (2000).

28 A number of observers argue that the IMF creates moral hazard even if its loans are repaid According to this view, the IMF enables the country’s government, captured by borrowers, to shift the losses from borrowers and lenders to the local taxpayer by providing a bridge loan.

29 For example, it is often argued that the IMF and the US Treasury bailed out Mexico in 1995 This is an improper use

of the term, unless clear evidence is supplied for the thesis that the o cial sector’s loans were unlikely to be repaid and that the official sector was just lucky.

30 For bank loans, a case can be made that foreign creditors do not only commit their own money, but also that of their home countries, since the lending banks may need to be rescued by taxpayer funds after losing money in crisis countries.

To a large extent, however, this problem is of concern primarily to the advanced countries’ supervisory authorities and deposit insurance funds There is not that much di erence between Credit Lyonnais’ investing in frail-looking companies and its lending to crisis-prone developing countries In both cases, the French taxpayers foot the bill.

31 In contrast, Burnside et al (2001) argue that it is government guarantees that make it optimal for banks to have an unhedged currency mismatch between their assets and their liabilities.

32 Some proponents of the panic view actually regard this lack of predictive power as a strength of the theory since the recent crises were poorly foreseen by observers On the other hand, the fundamental view and other views may be consistent with poorly foreseen crises as the countrywide exposures are hard to measure, as we will later argue Furthermore, one cannot exclude that advances in our theoretical and empirical knowledge of the functioning of this new breed of crises will make them easier to forecast It should also be noted that runs-based approaches may be combined with fundamentals-based theories Crises are then driven partly by weak fundamentals and partly by runs An illustration

of this mixed approach is provided by Chari–Kehoe (2001), in which foreign investors receive private signals about country strength and engage in (privately rational) herd behavior.

Outline of the Argument and Main Message

The discussion in Chapter 2 is shadowed by a lingering question: What are we trying todo? Preventing crises is not a goal in itself; after all, prohibiting foreign borrowingwould eliminate the threat of foreign debt crisis altogether! The issue therefore is, howdesirable are speci c policies when trying to accomplish a well-stated, unambiguousobjective?

In this respect, I am struck by the fact that proposals for a new internationalnancial architecture rarely formulate a clear objective function Or else, and almostequivalently, they o er a whole array of objectives: avoid nancial crises, resolve them

in an orderly manner, economize on taxpayer money, protect foreign investors, respectcountry sovereignty, limit output volatility, prevent contagion, facilitate the country’saccess to funds, promote long-term growth, and force structural reforms – to which can

be added the IMF’s traditional current account, international reserves, and in ationobjectives There are just too many goals, not to mention the fact that many of them arecon icting! Clearly, we need to go back to fundamentals and identify the marketfailures, if any, that underlie the existence of IFIs

Let us begin our theoretical treatment with the architecture of the argument andrelegate the details to the subsequent chapters

The problem of a standard borrower

Going back to basics leads us to revisit the obvious point that the unit of analysis is therelationship between a “borrower” (the debtor country) and a “lender” (theinternational financial community)

To secure funds, a borrower needs to put in place rules of behavior, that will applyboth before and after the uncertainty is resolved Financing arrangements have toresolve the basic problem of reassuring investors about the prospect of recouping theirinvestment In so doing, they need to strike a balance between the costs and bene ts ofalternative nancing and governance institutions (allocating control and cash owrights, de ning a monitoring structure, liquidity and risk management, bankruptcy andworkout procedures)

Investor protection bene ts the borrower more than the investors, as long as thelatter compete for the borrower’s business

Why is external borrowing different?

The need to put in place nancing arrangements that reassure investors about theprospect of recouping their investment exists at both the domestic and internationallevel What makes the latter special is the presence of a “third player”, the borrower’sgovernment, who shares with the private domestic borrowers the control rights notvested with investors, who is able to heavily in uence the return e ectively obtained bythe investors, and with whom investors do not contract Because the investors’ return is

a ected by the behavior of two agents, the borrower himself and its government, I call

this the dual-agency problem (Figure 4)

Figure 4

The government has both the incentives and the means not to fully protect the foreign

investors Governments do not derive any more utility from lowering the return onforeign investments than homeowners do from setting re to their house once insured.They do not per se enjoy devaluing the claims of foreigners and mostly do not do itwillingly Rather, they are less cautious than they should be at the margin; they will nottake any action that reduces the probability of a crisis if it entails a substantial politicalcost

Incentives to favor, at the margin, domestic constituencies over foreign investors areprovided either by democratic accountability, that induces governments to pander todomestic political constituencies, or by con icts of interest and capture, that inducepoliticians to favor speci c interest groups The means are the extensive control rightsheld by governments These are much more important in an international context than

domestically, since a myriad of government actions a ect the exchange rate and, moregenerally, the returns enjoyed by foreign investors.

The same conjunction of large discretionary powers and biased incentives engenders

a second key di erence between domestic corporate borrowing and international

borrowing: the common agency problem, which applies directly to sovereign borrowing,

and indirectly (through the dual agency problem) to private borrowing as well Whenseveral lenders lend to a single borrower, each lender takes no account of the impact ofits lending and associated contractual provisions on the other lenders This lendingexternality is present in corporate borrowing, with the presence of a large number ofequityholders, of bondholders, and of multiple securities But some of the responses tothis common agency problem that have been designed in the corporate context, have nonatural counterpart in the international context

Institutional and policy responses to market failure

Financial arrangements at the international level are improved if some institution (theIMF, say) acts as a delegated monitor on behalf of investors and is able to contract withthe government, thus “re-balancing” the dual- and the common-agency problems,ultimately to the benefit of the borrowing country

This perspective provides a conceptual framework in which to assess the currentproposals for redesigning the international nancial architecture and identi es themotivation for refocusing the IMF The framework internalizes the reasons why somefeatures of existing nancial arrangements are present in the rst place, thus avoidingthe risk of merely treating the symptoms, and considers the impact of privately optimalnancial arrangements on a government’s incentives and behavior The absence ofincentive in individual lending relationships to constrain government moral hazard has

a number of nonstandard, but intuitive implications

Liquidity and Risk-Management in a Closed Economy

A central question addressed in this book is the extent to which a country resembles or

di ers from an ordinary borrower To set the stage, we therefore need to recap the mainfeatures of nancing agreements My coverage of corporate nancing will be sketchyand highly selective, and will focus on the themes that are most relevant to thesubsequent chapters For obvious reasons, I will put special emphasis on control rightsand liquidity and risk management issues I will then discuss the notion and the role ofdomestic liquidity1 before turning to the international context in the next chapter

Corporate financing: key organizing principles

Corporate nance refers to the set of institutions and policies that make it credible thatthe suppliers of funds to a rm will recoup their investment, and thereby make itpossible for the borrower to have access to funds in the rst place Corporate nance istherefore about harnessing income that can be pledged to investors

From this perspective, all institutions and practices of corporate nancing, solvencyratios, liquidity requirements, lines of credit, duration analysis, currency matching,corporate governance, and bankruptcy processes, have a common rationale They all

re ect the existence of a wedge between the total value generated by the productiveactivity and the smaller amount that can be credibly promised to outside investors, what

we will call for lack of a better term “pledgeable income” This wedge between value and

pledgeable income is due to two factors or types of nonpledgeable income First, insiders

derive private bene ts from their position (employment, prestige, perks, etc.) thatcannot be appropriated by the rm’s nanciers Second, insiders may need to beprovided with financial incentives if their policy preferences are to be somewhat alignedwith those of the investors; without such nancial incentives, they may engage in petprojects and build empires, pay insu cient attention to subcontracting or labor costs,overlook their employees’ risk management procedures, allocate too little time to theirposition by overcommitting themselves with competing activities (boards of directors,political activism, investments in other ventures), enjoy excessive perks, favor theirfriends, or engage in a wide variety of self-dealing behaviors In a nutshell, and fromthe investors’ viewpoint, they may “misbehave”

That the outside investors cannot (and actually may not want to) appropriate thefull value is the essence of credit rationing – the fact that some socially desirableprojects do not receive funding from capital markets – but it has much broader

To have access to funds, rms need to take steps to reassure the investors as to theprospect of recouping their investment A rst point worth noting is that investorprotection is not so much about protecting investors as about bene tting borrowers.Investor protection is about enabling borrowers to have access to funds Indeed, forgiven market conditions and with competitive capital markets, a more credible promise

of returns to outside investors translates into a more intense competition to lend to theborrowers; investors are then willing to contribute more funds, lend at lower interestrates, demand less collateral, commit more liquidity, and take fewer control rights.Conversely, poor investor protection results at best in higher rates of interests andtougher conditions,3 and at worst in the absence of financing

This basic fact, predicted by theory and corroborated by empirical evidence, does notimply that investors have no incentive to lobby for increased protection Indeed, theyhave two reasons to do so First, increased protection bene ts their existing assets frompast investments Second, increased protection boosts new investment and therefore thedemand for savings Increased demand for savings raises equilibrium rates of return inthe capital market

Principle 1 (bene ciaries): With competitive capital markets, borrowers are the direct

bene ciaries of increased investor protection, as investors compete away the bene ts from protection Investors bene t indirectly to the extent that the associated increase in investment raises equilibrium rates of return.

For issues of interest, the steps taken to reassure investors are necessarily costly torms and reduce the net present value (NPV), or “value” for short from now on Astraightforward analogy is that of an entrepreneur pledging a family house with highsentimental value as collateral to a bank in order to be able to borrow Pledging thehouse as collateral boosts the income that is pledged to the investors, the pledgeableincome, but because it implies an ine cient transfer of property in the event theentrepreneur fails, the policy destroys value This ine ciency may be the cost to be paid

by the entrepreneur to have access to funding More generally, we can state:

Principle 2 (pledgeable income – value trade-o ): Borrowers face a trade-o between

pledgeable income and value.

While the costly-collateral-pledging example provides the intuition, let us give a fewother applications of this principle:

(a) Corporate governance.

Consider, rst, the institutions of corporate governance These institutions involvesubstantial transaction costs, delays and lack of exibility in managerial decision-making, disclosure of information that may be useful to competitors, and managerialdistraction away from core tasks Corporate governance is entirely about raising thepledgeable income The fact that borrowers are willing to incur the corresponding costs

is clear evidence of the pledgeable income – value trade-off